Simple Reinforcement Learning: Temporal Difference Learning
So recently I’ve been doing a lot of reading on reinforcement learning and watching David Silver’s Introduction to Reinforcement Learning video series, which by the way are phenomenal and I highly recommend them! Coming from a traditional statistics and machine learning background, in terms of both grad school and work projects, these topics were somewhat new to me. So for my own personal learning and to share that learning with those interested I thought I’d archive it through a Medium post while trying to make these concepts as simple as possible to understand.
Why Reinforcement Learning
Why would we want to use reinforcement learning versus some other supervised learning approach? To answer that I came up with this timely example:
Let’s say you’re a big fan of the NFL and you want to predict number of wins for the upcoming season. At the beginning of the season you may look at variables such as: past season wins, number of players injured, forecasted weather conditions, number of 1st year starting players, etc. You then fit a model and predict a 9-win season. You have a good team, but an unproven and new quarterback (only 1 game less compared to last year). After the first 5 games your team has a record of 5–0 and your new quarterback is already being hailed as a future…
